Power Increases the Socially Toxic Component of Narcissism Among Individuals with High Baseline Testosterone

Journal of Experimental Psychology: General

BRIEF REPORT

Power Increases the Socially Toxic Component of Narcissism Among Individuals With High Baseline Testosterone

Nicole L. Mead Roy F. Baumeister University of Melbourne University of Queensland and Florida State University

Anika Stuppy Kathleen D. Vohs Erasmus University Rotterdam University of Minnesota

The corrosive effects of power have been noted for centuries, but the self-related changes responsible for those effects have remained somewhat elusive. Narcissists tend to rise to—and abuse—positions of power, so we considered the possibility that positions of power may corrupt because they inflate narcissism. Two pathways were considered: Powerholders abuse their power because having power over others makes them feel superior (grandiosity pathway) or deserving of special treatment (entitlement pathway). Supporting the entitlement pathway, assigning participants to a position of power (vs. equal control) over a group task increased scores on the Exploitative/Entitlement component of narcissism among those with high baseline testosterone. What is more, heightened Exploitative/Entitlement scores among high-testosterone participants endowed with power (vs. equal control) statistically explained amplified self-reported willingness to misuse their power (e.g., taking fringe benefits as extra compen- sation). The grandiosity pathway was not well supported. The Superiority/Arrogance, Self-Absorption/ Self-Admiration, and Leadership/Authority facets of narcissism did not change as a function of the power manipulation and testosterone levels. Taken together, these results suggest that people with high (but not low) testosterone may be inclined to misuse their power because having power over others makes them feel entitled to special treatment. This work identifies testosterone as a characteristic that contributes to the development of the socially toxic component of narcissism (Exploitative/Entitlement). It points to the possibility that structural positions of power and individual differences in narcissism may be mutually reinforcing, suggesting a vicious cycle with personal, relational, and societal implications.

Keywords: narcissism, power, testosterone, corruption, entitlement

Supplemental materials: http://dx.doi.org/10.1037/xge0000427.supp

For centuries power has been viewed as a corrupting force. influence of power have remained somewhat elusive. The so- Despite the recent surge in psychological research examining cially toxic behaviors of the powerful resemble those of nar- the cognitive, motivational, and behavioral consequences of cissists, so we investigated the possibility that social power power, the self-related changes that explain the corrupting increases narcissism. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Nicole L. Mead, Department of Management and Marketing, Faculty of (SSHRC; Nicole L. Mead) and the Erasmus Research Institute of Business and Economics, University of Melbourne; Roy F. Baumeister, Management (ERIM; Nicole L. Mead and Anika Stuppy). The content School of Psychology, University of Queensland, and Department of is solely the responsibility of the authors and does not necessarily Psychology, Florida State University; Anika Stuppy, Department of Mar- reflect the official views of the SSHRC or ERIM. We thank Cristina Ion keting Management, Rotterdam School of Management, Erasmus Univer- and Victoria Cazacu for assistance with data collection and Bram Van sity Rotterdam; Kathleen D. Vohs, Department of Marketing, Carlson den Bergh and Jon Maner for feedback on an earlier version of the School of Management, University of Minnesota. article. The experiment reported in this article has not been shared at profes- Correspondence concerning this article should be addressed to Nicole sional meetings, nor have the data been disseminated. L. Mead, Department of Management and Marketing, Faculty of Busi- This work benefited from the support of a doctoral fellowship from ness and Economics, University of Melbourne, Level 10, 198 Berkeley the Social Sciences and Humanities Research Council of Canada Street, 3010 Victoria, Australia. E-mail: [email protected]

591 592 MEAD, BAUMEISTER, STUPPY, AND VOHS

Two Possible Power-to-Corruption Pathways power inflates the Exploitative/Entitlement component of narcissism among those with high testosterone. An experiment tested Narcissism consists of two separate components: entitlement these hypotheses by measuring individual differences in circulat- and grandiosity (Brown, Budzek, & Tamborski, 2009). Entitlement ing testosterone and manipulating whether participants believed is the desire to be recognized and treated as special by others. they had power (vs. equal control) over a group task. Grandiosity is relatively more intrapersonal and is characterized by an arrogant sense of self-importance. Method The maladaptive behaviors and interpersonal problems that characterize abusive powerholders, such as aggression, cheating, Participants and Procedure and counterproductive workplace behaviors, have been linked with the Exploitative/Entitlement component of narcissism (for a re- Previous studies examining the interaction between manipulated view, see Grijalva et al., 2015). Hence, increased entitlement and hierarchical position and measured testosterone levels have used exploitation seemed a viable explanation for the corrupting influ- inconsistent sample sizes. We therefore used a rule of thumb of at ence of power. We label this the entitlement pathway. least 50 participants per cell (Simmons, Nelson, & Simonsohn, An alternative possibility is that power corrupts because it 2013). By the end of recruitment (the day we reached 200 partic- makes people think they are better than others. We label this the ipants), 206 participants (98 female) completed the study in ex- grandiosity pathway. Conceptually, this pathway is similar to the change for monetary payment. The study was approved by the theory that power corrupts through inflated self-esteem (Kipnis, school’s institutional review board. 1972) because the grandiosity components of narcissism are cor- To increase the believability of the role-play power manipula- related with self-esteem (whereas the Exploitative/Entitlement tion, we arranged for participants to arrive in groups for a team- component is not; Emmons, 1984, 1987; Watson & Biderman, dynamics study. The experimenter brought the group to a room 1993). that was labeled group room and was set up to facilitate a group The evidence for the link between power and self-esteem has interaction. Participants were told they would complete a team task been mixed, however. Recalling a time of power (vs. recalling together there. Then they were led to individual cubicles, where yesterday’s activities) increased self-esteem (Fast, Gruenfeld, they were asked to complete initial measures. In reality, those tasks Sivanathan, & Galinsky, 2009). Yet giving people actual power constituted the study procedures. All study materials are public (vs. equal control) over a group task did not change self-esteem (Mead, 2017). (Kipnis, 1972; Wojciszke & Struzynska-Kujalowicz, 2007). Thus, self-esteem did not explain the corrupting influence of power Testosterone (Kipnis, 1972). Given that the Exploitative/Entitlement component of narcis- Participants completed a testosterone briefing session 24 hr sism has been linked with a wide range of socially toxic behaviors, before the main experiment. The session aimed to ensure that we favored the entitlement pathway. Nevertheless, we examined participants would provide a clean saliva sample during the main both possible pathways by examining how power changes scores experiment (see the online supplemental materials). on the narcissism components that capture grandiosity and entitle- To minimize circadian fluctuations in testosterone, we con- ment. ducted the experiment between 12 (noon) and 3 p.m. (Mehta, Jones, & Josephs, 2008). Saliva was collected before the power Testosterone manipulation. Participants drooled 1.5 ml of saliva through a straw into a sterile polypropylene SaliCap tube. After collection, saliva Although the notion that social power corrupts has anecdotal samples were frozen at Ϫ20 °C. Samples were shipped on dry ice and scientific support (e.g., Kipnis, 1972), not all people misuse to Clemens Kirschbaum’s laboratory. Sampling tubes were centri- their position of power. We focused on testosterone as an attribute fuged at 10,000 rpm for 5 min. Salivary testosterone concentrations that may predispose people to the corrupting influence of structural were measured using commercially available chemiluminescence- power. Leaders with high testosterone were prone to use their immuno-assays with high sensitivity (lower limit of detection ϭ position of power to improve their own outcomes at the expense of 1.8 pg/ml; analytical range ϭ 1.8Ϫ760 pg/ml). Six participants This document is copyrighted by the American Psychological Association or one of its allied publishers. others (Bendahan, Zehnder, Pralong, & Antonakis, 2015). If nar- provided unusable saliva samples. The inter- and intraassay coef- This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. cissism is the missing link between power and corruption, then ficients were 5.68% and 5.39%, respectively. Men (M ϭ 101.38 gains in power should have the strongest effect on narcissism pg/mL, SD ϭ 54.91) registered testosterone levels that were ap- among those with high testosterone. proximately three times higher than women’s (M ϭ 30.26 pg/mL, Being endowed with power may increase entitlement and ex- SD ϭ 21.23), t(190) ϭ 11.647, p Ͻ .0001, which is consistent with ploitation among high-testosterone individuals. Testosterone is guidelines for determining that testosterone analyses were specific positively associated with dominant behavior that is intended to to testosterone (Schultheiss & Stanton, 2009). achieve or maintain high social rank (Archer, 2006; Mazur & Power manipulation. We used a previously validated power Booth, 1998). Those highly motivated to dominate others strive to manipulation that gives participants asymmetric (vs. equal) control retain their positions of power, even at the expense of group over a group task and rewards (Case & Maner, 2014; Maner & interests (Maner & Mead, 2010; Mead & Maner, 2012a, 2012b). Mead, 2010; Mead & Maner, 2012a). Participants completed two Garnering special treatment and manipulating others for one’s measures that were said to measure leadership abilities. We told all purposes may help to maintain or even increase a power differen- participants they had achieved the highest combined score on the tial between the self and others. We therefore predicted that social leadership measures. Therefore, any changes in narcissism could POWER AND NARCISSISM 593

be attributed to differences in interpersonal power (described in the be linked to androgen-receptor genes (Voracek, 2014) or corre- next paragraph) rather than performance feedback. lated with testosterone (Hönekopp, Bartholdt, Beier, & Liebert, Participants randomly assigned to the power condition were 2007). informed they would be the “Boss” of the group task because of We measured digit ratio. Images of participants’ right hands their top scores on the leadership tasks. As Boss, they would were acquired via a flatbed scanner. Digit length was measured instruct the other group members (“subordinates”) about how to from the metacarpophalangeal crease to the tip of the finger. Eight perform the team task, evaluate their subordinates throughout the participants provided unusable (blurry) hand scans, resulting in a task, and decide how to distribute monetary rewards that would be final sample of 192 participants (92 female; 101 equal-control ϭ earned during the group task. In contrast, participants in the condition; Mage 21.97 years). Digit ratio was calculated by equal-control condition were told that all group members had dividing the length of the second digit by the length of the fourth ϭ ϭ ϭ ϭ equal control over the group task and that the monetary rewards digit (Mmen .95, SD .03; Mwomen .96, SD .03). Consis- earned during the group task would be divided equally among tent with meta-analytic conclusions (Hönekopp et al., 2007), digit group members. As a manipulation check, participants indicated ratio and testosterone did not covary among men, r(99) ϭϪ.021, how powerful they felt on a scale of 0 (not at all) to 100 (very p ϭ .839, or women, r(91) ϭ .050, p ϭ .638. much so; M ϭ 58.87, SD ϭ 24.09). Ostensibly because the group room was not yet available, par- Results ticipants were asked to complete additional measures while they waited to start the group task. In reality, those were the dependent measures. The cover story was given to encourage continued Predictive Model feelings of interpersonal power during the completion of the outcome variables. In a funnel debriefing, only two participants To account for gender differences in the predictors testosterone expressed substantial suspicion about the manipulation or group and 2D:4D, we standardized those variables within gender (Maner, task, suggesting the cover story was successful. Miller, Schmidt, & Eckel, 2008). Preliminary analyses revealed Narcissism. Narcissism was assessed with the commonly that power condition did not interact with gender or digit ratio; used 40-item Narcissistic Personality Inventory (NPI; Raskin & therefore, those are not discussed further. Full results are reported Terry, 1988). Participants choose between two options, with one in the online supplemental materials. option being more narcissistic than the other. The instructions were Men (vs. women) have been found to be more narcissistic modified by informing participants to respond in terms of their (Grijalva et al., 2015) and have reported being more willing to momentary feelings (Giacomin & Jordan, 2014). misuse their power (Lee-Chai et al., 2001), so we controlled for We computed scores for the four factors identified by Emmons gender to isolate the effects of interest. Excluding gender as a (1984, 1987). Exploitative/Entitlement (e.g., “I insist on getting the covariate did not change the interpretation of the results (see the respect that is due to me”; M ϭ 2.53, SD ϭ 1.86; ␣ϭ.59) was online supplemental materials). The predictive model for main-text used to test the entitlement pathway. Self-absorption/Self- analyses was as follows: power condition (centered), testosterone admiration (e.g., “I am an extraordinary person”; M ϭ 3.66, SD ϭ (standardized within gender), participant gender (centered), and 2.10; ␣ϭ.66) and Superiority/Arrogance (e.g., “I can make the theoretically relevant interaction (Power Condition ϫ Testos- anybody believe anything”; M ϭ 2.74, SD ϭ 1.70; ␣ϭ.54) were terone Levels). clear tests of the grandiosity pathway. Leadership/Authority (e.g., ϭ ϭ ␣ϭ “I see myself as a good leader”; M 4.31, SD 2.42; .77) Power Manipulation Check was not a conceptually clear test of grandiosity, but its positive correlation with self-esteem (e.g., Emmons, 1984) suggested it Endowing participants with power (vs. equal control) over the may capture positive self-views. Thus, the Leadership/Authority group task heightened self-reported feelings of power (␤ϭ.157), factor was computed and examined as a potential test of the t(187) ϭ 2.178, p ϭ .031, partial r ϭ .157; participants in the grandiosity pathway. power condition felt more powerful (M ϭ 62.84, SD ϭ 21.74) than Corruption. Willingness to abuse one’s power was assessed did participants in the equal-control condition (M ϭ 55.30, SD ϭ by summing responses to the 18-item Misuse of Power scale 25.61). There were no other significant predictors (ps Ͼ .295). This document is copyrighted by the American Psychological Association or one of its allied publishers. (Lee-Chai, Chen, & Chartrand, 2001). This scale captures willing- Hence, the manipulation worked regardless of testosterone levels. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. ness to use one’s power to improve one’s outcomes at the expense of others (e.g., “There is nothing wrong with occasionally taking Entitlement credit for one of your subordinates’ ideas”), rated on a scale of 1 (strongly disagree)to9(strongly agree; M ϭ 71.11, SD ϭ 21.09; Full results are reported in Table 1. The predicted interaction ␣ϭ.81). It predicts willingness to misuse power in a variety of between power condition and testosterone levels was not signifi- specific situations (e.g., accept a bribe) and has predictive value cant (␤ϭ.124), t(187) ϭ 1.701, p ϭ .091, partial r ϭ .123; see beyond the constructs of dominance and exploitation (Lee-Chai et Figure 1), but we proceeded with a priori hypothesis testing al., 2001). (Iacobucci, 2001; Keppel & Wickens, 2004; Winer, Brown, & Digit ratio. Second digit (index) to fourth digit (ring) ratio Michels, 1991). Supporting the entitlement pathway, endowing (2D:4D) has been used as a proxy for testosterone levels (e.g., participants with power (vs. equal control) over the group task in- Ronay & von Hippel, 2010) because it has been argued that 2D:4D creased Exploitative/Entitlement scores among high-testosterone (ϩ1 reflects exposure to androgens in the womb (Manning, 2002). SD from the mean; ␤ϭ.233), t(187) ϭ 2.278, p ϭ .024, partial However, recent meta-analyses concluded that digit ratio may not r ϭ .164, but not low-testosterone (Ϫ1 SD from the mean; 594 MEAD, BAUMEISTER, STUPPY, AND VOHS

Table 1 Results of Multiple Regression Analyses Predicting Narcissism Facets

Exploitative/Entitlement Self-absorption/Self-admiration Superiority/Arrogance Outcome variable ␤ tppartial r ␤ tppartial r ␤ tppartial r

Power condition .109 1.525 .129 .111 .102 1.410 .160 .103 .023 .326 .745 .024 Testosterone .025 .344 .732 .025 .019 .260 .795 .019 .045 .630 .529 .046 Gender .123 1.713 .088 .124 .138 1.918 .057 .139 .229 3.238 .001 .230 Power ϫ Testosterone .124 1.701 .091 .123 .050 .688 .492 .050 .115 1.604 .110 .117 Power condition at ϩ1 SD testosterone .233 2.278 .024 .164 .152 1.478 .141 .107 .138 1.368 .173 .100 Power condition at Ϫ1 SD testosterone Ϫ.014 .137 .891 Ϫ.010 .052 .503 .616 .037 Ϫ.092 .913 .363 Ϫ.067

␤ϭϪ.014), t(187) ϭ .137, p ϭ .891, partial r ϭϪ.010, partic- and testosterone levels (␤ϭ.149), t(187) ϭ 2.081, p ϭ .039, ipants. partial r ϭ .150 (see Figure 2). Consistent with predictions, the power manipulation increased self-reported misuse of power Grandiosity among high-testosterone (ϩ1 SD; ␤ϭ.267), t(187) ϭ 2.666, p ϭ .008, partial r ϭ .191 (see Figure 2), but not low-testosterone (Ϫ1 Regressing the Superiority/Arrogance and Self-absorption/Self- SD; ␤ϭϪ.029), t(187) ϭ .290, p ϭ .772, partial r ϭϪ.021, Admiration factors on the main predictive model revealed main participants. effects of only gender. Full results are reported in Table 1. The main model revealed a main effect of gender (␤ϭ.218), t(187) ϭ 3.106, p ϭ .002, partial r ϭ .221. Power condition (␤ϭ Leadership/Authority .119), t(187) ϭ 1.693, p ϭ .092, partial r ϭ .123, and testosterone levels (␤ϭ.018), t(187) ϭ .245, p ϭ .806, partial r ϭ .018, were Regressing Leadership/Authority scores on the main predictive not significant predictors. model revealed a main effect of gender (␤ϭ.260), t(187) ϭ 3.689, p Ͻ .0001, partial r ϭ .260, whereby male participants (M ϭ 4.90, SD ϭ 2.27) endorsed more narcissistic Leadership/ Process Analyses ϭ Authority statements than did female participants (M 3.66, We tested which narcissism facet explained the interactive ef- ϭ ␤ϭ SD 2.43). Power condition was not a significant predictor ( fect of power condition and testosterone levels on misuse of power ϭ ϭ .030), t(187) .423, p .673, nor was baseline testosterone scores. We estimated a moderated-mediation model with 5,000 ␤ϭϪ ϭ ϭ ( .060), t(187) .839, p .403. The interaction between bias-corrected bootstrap samples (Hayes, 2013; Model 8), control- ␤ϭ power condition and testosterone levels was not significant ( ling for gender (excluding gender did not change the conclusions ϭ ϭ .078), t(187) 1.090, p .277. The simple effect of the power of the results; see the online supplemental materials). manipulation (vs. equal control) was not significant among high- Supporting theorizing, the indirect effect of power condition on ␤ϭ ϭ ϭ testosterone ( .108), t(187) 1.070, p .286, or low- misuse of power scores via Exploitative/Entitlement scores was ␤ϭϪ ϭ ϭ testosterone ( .048), t(187) .478, p .633, participants. significant for high-testosterone (ϩ1 SD; 95% confidence interval [CI] [.570, 7.806]) but not low-testosterone (Ϫ1 SD; 95% CI Misuse of Power [Ϫ3.640, 2.900]) participants. Regressing Misuse of Power scores on the main predictive model revealed the predicted interaction between power condition This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Figure 1. Endowing participants with power (vs. equal control) over a Figure 2. Being assigned to a position of power (vs. equal control) over group task increased scores on the Exploitative/Entitlement component of a group task heightened reported misuse of power among high-testosterone narcissism among participants with high (ϩ1 SD) but not low (Ϫ1 SD) (ϩ1 SD) but not low-testosterone (Ϫ1 SD) participants. Unstandardized testosterone levels. Unstandardized regression coefficients compare the regression coefficients compare the power condition to the equal-control .p ϭ .008 ءء .p ϭ .024. condition ء .power condition to the equal-control condition POWER AND NARCISSISM 595

This was not the case for Superiority/Arrogance scores: high- terone may predispose people to the corrupting influence of power testosterone (ϩ1 SD; 95% CI [Ϫ.661, 1.870]) and low- (Bendahan et al., 2015) and has been theoretically linked to nar- testosterone (Ϫ1 SD; 95% CI [Ϫ2.080, .453]) participants; Self- cissism (Grijalva et al., 2015), so we investigated its moderating absorption/Self-admiration scores: high-testosterone (ϩ1 SD; role in the current work. 95% CI [Ϫ.694, 1.999] and low-testosterone (Ϫ1 SD; 95% CI [Ϫ.439, 1.536]) participants; or Leadership/Authority scores: high- Conclusion testosterone (ϩ1 SD; 95% CI [Ϫ2.111, .316]) and low- testosterone (Ϫ1 SD; 95% CI [Ϫ.460, 2.010]) participants. Hence, Power is an essential component of social life. Although the the effect of the power manipulation on self-reported misuse of corrupting nature of power long has been noted, the self-related power was mediated by increased Exploitative/Entitlement scores processes responsible have remained an enigma. The current find- among participants with relatively high testosterone levels. ings suggest that entitlement may be a missing piece of the puzzle. They indicate that although power does not turn everyone into corruptive tyrants, it does have the most pernicious consequences Discussion when it gets into the hands of those who want it the most. We tested two self-related pathways that may help explain the corrupting influence of power. The first was that power corrupts References because it leads people to think they are better than others (gran- Archer, J. (2006). Testosterone and human aggression: An evaluation of diosity pathway). The second was that power corrupts because it the challenge hypothesis. Neuroscience and Biobehavioral Reviews, 30, makes people feel entitled to special treatment (entitlement path- 319–345. http://dx.doi.org/10.1016/j.neubiorev.2004.12.007 way). Those who enjoy power try to keep it, even at the expense Bendahan, S., Zehnder, C., Pralong, F. P., & Antonakis, J. (2015). Leader of others (Maner & Mead, 2010; Mead & Maner, 2012a), so we corruption depends on power and testosterone. Leadership Quarterly, predicted that power would be especially likely to foster entitled 26, 101–122. http://dx.doi.org/10.1016/j.leaqua.2014.07.010 self-views among those with high testosterone. Brown, R. P., Budzek, K., & Tamborski, M. (2009). On the meaning and measure of narcissism. Personality and Social Psychology Bulletin, 35, Supporting the entitlement pathway, giving participants power 951–964. http://dx.doi.org/10.1177/0146167209335461 (vs. equal control) over a group task increased scores on the Case, C. R., & Maner, J. K. (2014). Divide and conquer: When and why Exploitative/Entitlement component of narcissism—but only leaders undermine the cohesive fabric of their group. Journal of Per- among those with high testosterone. Furthermore, the power ma- sonality and Social Psychology, 107, 1033–1050. http://dx.doi.org/10 nipulation increased stated willingness to misuse power among .1037/a0038201 high- but not low-testosterone participants. Inflated Exploitative/ Emmons, R. A. (1984). Factor analysis and construct validity of the Entitlement scores among high-testosterone participants in the Narcissistic Personality Inventory. Journal of Personality Assessment, power (vs. equal-control) condition statistically accounted for 48, 291–300. http://dx.doi.org/10.1207/s15327752jpa4803_11 heightened self-reported willingness to misuse power. Future re- Emmons, R. A. (1987). Narcissism: Theory and measurement. Journal of search could examine whether these results depend on receiving Personality and Social Psychology, 52, 11–17. http://dx.doi.org/10 .1037/0022-3514.52.1.11 social power and positive feedback, given that our manipulation Fast, N. J., Gruenfeld, D. H., Sivanathan, N., & Galinsky, A. D. (2009). endowed participants with both. Illusory control: A generative force behind power’s far-reaching effects. Consistent with previous work (Kipnis, 1972; Wojciszke & Psychological Science, 20, 502–508. http://dx.doi.org/10.1111/j.1467- Struzynska-Kujalowicz, 2007), giving participants actual social 9280.2009.02311.x power (vs. equal control) over a group task did not inflate self- Giacomin, M., & Jordan, C. H. (2014). Down-regulating narcissistic ten- views. The empirical divergence of Exploitative/Entitlement from dencies: Communal focus reduces state narcissism. Personality and the other narcissism factors is consistent with previous work Social Psychology Bulletin, 40, 488–500. http://dx.doi.org/10.1177/ (Emmons, 1984, 1987; Watson & Biderman, 1993). However, the 0146167213516635 manipulation check suggested that our power manipulation was Grijalva, E., Newman, D. A., Tay, L., Donnellan, M. B., Harms, P. D., relatively weak. A stronger power manipulation may produce Robins, R. W., & Yan, T. (2015). Gender differences in narcissism: A meta-analytic review. Psychological Bulletin, 141, 261–310. http://dx different results. .doi.org/10.1037/a0038231 Digit ratio did not moderate the effects of power. Although This document is copyrighted by the American Psychological Association or one of its allied publishers. Hayes, A. F. (2013). Introduction to mediation, moderation, and condi- 2D:4D has been used as an indirect measure of in utero exposure

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. tional process analysis: A regression-based approach. New York, NY: to androgens, it has been suggested that the hormonal processes Guilford Press. that give rise to 2D:4D may not be androgen-related (Voracek, Hönekopp, J., Bartholdt, L., Beier, L., & Liebert, A. (2007). Second to 2014). Future research should continue to uncover the similarities fourth digit length ratio (2D:4D) and adult sex hormone levels: New data and differences between 2D:4D and testosterone to better under- and a meta-analytic review. Psychoneuroendocrinology, 32, 313–321. stand what overlapping psychological correlates they may have. http://dx.doi.org/10.1016/j.psyneuen.2007.01.007 Iacobucci, D. (2001). Quantitative marketing research. In D. Iacobucci (Ed.), Kellogg on marketing (pp. 195–211). New York, NY: Wiley. Broader Context Keppel, G., & Wickens, T. D. (2004). Design and analysis: A researcher’s handbook (4th ed.). Upper Saddle River, NJ: Pearson Prentice Hall. This article originated from Nicole L. Mead’s interest in why Kipnis, D. (1972). Does power corrupt? Journal of Personality and Social people misuse their power. In early discussions, Nicole L. Mead Psychology, 24, 33–41. http://dx.doi.org/10.1037/h0033390 and Roy F. Baumeister noted the striking similarity between the Lee-Chai, A. Y., Chen, S., & Chartrand, T. L. (2001). From Moses to conduct of the powerful and the narcissistic, leading to the basic Marcos. Individual differences in the use and abuse of power. In A. Y. hypothesis of this work. Previous work has suggested that testos- Lee-Chai & J. A. Bargh (Eds.), The use and abuse of power: Multiple 596 MEAD, BAUMEISTER, STUPPY, AND VOHS

perspectives on the causes of corruption (pp. 57–74). New York, NY: Raskin, R., & Terry, H. (1988). A principal-components analysis of the Taylor & Francis. Narcissistic Personality Inventory and further evidence of its construct Maner, J. K., & Mead, N. L. (2010). The essential tension between validity. Journal of Personality and Social Psychology, 54, 890–902. leadership and power: When leaders sacrifice group goals for the sake of http://dx.doi.org/10.1037/0022-3514.54.5.890 self-interest. Journal of Personality and Social Psychology, 99, 482– Ronay, R., & von Hippel, W. (2010). Power, testosterone, and risk-taking. 497. http://dx.doi.org/10.1037/a0018559 Journal of Behavioral Decision Making, 23, 473–482. http://dx.doi.org/ Maner, J. K., Miller, S. L., Schmidt, N. B., & Eckel, L. A. (2008). 10.1002/bdm.671 Submitting to defeat: Social anxiety, dominance threat, and decrements Schultheiss, O. C., & Stanton, S. J. (2009). Assessment of salivary hor- in testosterone. Psychological Science, 19, 764–768. http://dx.doi.org/ mones. In E. Harmon-Jones & J. S. Beer (Eds.), Methods in the neuro- 10.1111/j.1467-9280.2008.02154.x biology of social and personality psychology (pp. 17–44). New York, Manning, J. T. (2002). Digit ratio: A pointer to fertility, behavior, and NY: Guilford Press. Life after health. New Brunswick, NJ: Rutgers University Press. Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2013, January). p-hacking. Meeting of the Society for Personality and Social Psychol- Mazur, A., & Booth, A. (1998). Testosterone and dominance in men. ogy, New Orleans, LA. http://dx.doi.org/10.2139/ssrn.2205186 Behavioral and Brain Sciences, 21, 353–363. http://dx.doi.org/10.1017/ Voracek, M. (2014). No effects of androgen receptor gene CAG and GGC S0140525X98001228 repeat polymorphisms on digit ratio (2D:4D): A comprehensive meta- Mead, N. L. (2017, October 13). Power-Narcissim-Corruption. http://dx analysis and critical evaluation of research. Evolution and Human Be- .doi.org/10.17605/OSF.IO/7E3C6 havior, 35, 430–437. http://dx.doi.org/10.1016/j.evolhumbehav.2014.05 Mead, N. L., & Maner, J. K. (2012a). On keeping your enemies close: .009 Powerful leaders seek proximity to ingroup power threats. Journal of Watson, P. J., & Biderman, M. D. (1993). Narcissistic Personality Inven- Personality and Social Psychology, 102, 576–591. http://dx.doi.org/10 tory factors, splitting, and self-consciousness. Journal of Personality .1037/a0025755 Assessment, 61, 41–57. http://dx.doi.org/10.1207/s15327752jpa6101_4 Mead, N. L., & Maner, J. K. (2012b). When me versus you becomes us Winer, B. J., Brown, D. R., & Michels, K. M. (1991). Statistical principles versus them: How intergroup competition shapes ingroup psychology. in experimental design (3rd ed.). New York, NY: McGraw-Hill. Social & Personality Psychology Compass, 6, 566–574. https://dx.doi Wojciszke, B., & Struzynska-Kujalowicz, A. (2007). Power influences .org/10.1111/j.1751-9004.2012.00447.x self–esteem. Social Cognition, 25, 472–494. http://dx.doi.org/10.1521/ Mehta, P. H., Jones, A. C., & Josephs, R. A. (2008). The social endocri- soco.2007.25.4.472 nology of dominance: Basal testosterone predicts cortisol changes and behavior following victory and defeat. Journal of Personality and Social Received March 15, 2017 Psychology, 94, 1078–1093. http://dx.doi.org/10.1037/0022-3514.94.6 Revision received January 19, 2018 .1078 Accepted February 15, 2018 Ⅲ

E-Mail Notification of Your Latest Issue Online!

Would you like to know when the next issue of your favorite APA journal will be available online? This service is now available to you. Sign up at https://my.apa.org/portal/alerts/ and you will This document is copyrighted by the American Psychological Association or one of its allied publishers. be notified by e-mail when issues of interest to you become available! This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.